Evidence · Public register

See what was tested. See what changed.

Explore the task, permitted inputs, comparison, accuracy and timing behind each reported result. Completed, partial and pending work remain distinguishable.

Illustrative optical loupe revealing a precise molecular lattice
Illustrative evidence-inspection visual · not evidence

BGC87 · Public summary · 28 September 2026

Scientific answers. Time to result.

87 planned cases across 13 field groupings. LINUR Engine, Opus, Sol and Astra compared through reference agreement, recorded delivery time and observed repeatability—three distinct measures.

35 shared complete-output timing cases. Median recorded intervals: LINUR 0.170 s; Opus 338.60 s; Sol 203.61 s; Astra 226.13 s.

Different timing boundaries. LINUR measures child dispatch-to-return, excluding wrapper preparation and packaging. AI sessions include tools from task receipt to answer-file finalization. This is not a like-for-like or common-hardware speed comparison.

Scoped computational comparison. Missing outputs and historical protocol decisions remain visible; reference agreement is not independent experimental validation.

Explore the public results ↗

Read the 17-page executive report ↗

Go deeper under NDA.

Detailed case records, sensitive references and supporting evidence are held in a separate confidential technical annex. Access requires an NDA and LINUR approval.

Request technical annex under NDA →

Latest comparison · 28 September 2026

Explore the new results by field.

87 planned cases across 13 field groupings. The report retains 492 trial records, including repeat and recovery records—not 492 distinct scientific problems. Reference agreement, delivery time and repeatability are assessed separately.

Browse the tested questions below. The earlier 14 application studies remain a separate supporting collection; their counts and evidence classifications are not combined with this campaign.

BGC87 · 8 planned tasks · Scoped computational comparison

Calibration

Quantum many-body reference; Main-group thermochemistry; Noncovalent interactions; Solid-state reproducibility; Lattice dynamics & thermal transport; Criticality & neutron transport; Liquid water structure & dynamics

Recorded delivery time

LINUR
0.326 s
Opus
202.02 s
Sol
225.47 s
Astra
196.73 s

1 matched timed tasks. Median seconds on shared complete-output cases.

LINUR: child dispatch-to-return, excluding wrapper preparation and packaging. AI: task receipt to answer-file finalization, including tools. Different boundaries—not like-for-like or hardware-normalized speed.

Reference agreement and scope

Linur: 4.30735e-4 relative-difference ratio.

Opus: 0 relative-difference ratio.

Sol: 0 relative-difference ratio.

Astra: 0 relative-difference ratio.

4 cases / 51 matched scalar outputs.

Relative difference is not an accuracy percentage. Scalar summaries are medians within, then across cases; medians can hide individual errors. Missing shared evidence is not a zero error.

View the field charts ↗
BGC87 · 6 planned tasks · Scoped computational comparison

Batteries

Lithium-metal & solid-state batteries; Solid-state electrolytes; Cathode degradation & thermal runaway; Fast charging & lithium plating

Recorded delivery time

LINUR
0.306 s
Opus
360.25 s
Sol
324.85 s
Astra
160.48 s

2 matched timed tasks. Median seconds on shared complete-output cases.

LINUR: child dispatch-to-return, excluding wrapper preparation and packaging. AI: task receipt to answer-file finalization, including tools. Different boundaries—not like-for-like or hardware-normalized speed.

Reference agreement and scope

Linur: 0.0739431 relative-difference ratio.

Opus: 0.0478905 relative-difference ratio.

Sol: 0.0610685 relative-difference ratio.

Astra: 0.0545425 relative-difference ratio.

6 cases / 218 matched scalar outputs.

Relative difference is not an accuracy percentage. Scalar summaries are medians within, then across cases; medians can hide individual errors. Missing shared evidence is not a zero error.

View the field charts ↗
BGC87 · 5 planned tasks · Scoped computational comparison

2D materials

Twistronics & quantum materials; 2D semiconductors for quantum technology; Graphene-based sensors

Recorded delivery time

LINUR
0.298 s
Opus
589.04 s
Sol
243.38 s
Astra
226.13 s

5 matched timed tasks. Median seconds on shared complete-output cases.

LINUR: child dispatch-to-return, excluding wrapper preparation and packaging. AI: task receipt to answer-file finalization, including tools. Different boundaries—not like-for-like or hardware-normalized speed.

Reference agreement and scope

Linur: 0 relative-difference ratio.

Opus: 0 relative-difference ratio.

Sol: 0 relative-difference ratio.

Astra: 0 relative-difference ratio.

5 cases / 43 matched scalar outputs.

Relative difference is not an accuracy percentage. Scalar summaries are medians within, then across cases; medians can hide individual errors. Missing shared evidence is not a zero error.

View the field charts ↗
BGC87 · 8 planned tasks · Scoped computational comparison

Renewables

Perovskite photovoltaics; Electrocatalysis; Thermoelectrics; Hydrogen storage

Recorded delivery time

LINUR
0.181 s
Opus
302.29 s
Sol
149.61 s
Astra
256.20 s

8 matched timed tasks. Median seconds on shared complete-output cases.

LINUR: child dispatch-to-return, excluding wrapper preparation and packaging. AI: task receipt to answer-file finalization, including tools. Different boundaries—not like-for-like or hardware-normalized speed.

Reference agreement and scope

Linur: 9.1261e-7 relative-difference ratio.

Opus: 9.1261e-7 relative-difference ratio.

Sol: 9.1261e-7 relative-difference ratio.

Astra: 9.1261e-7 relative-difference ratio.

8 cases / 65 matched scalar outputs.

Relative difference is not an accuracy percentage. Scalar summaries are medians within, then across cases; medians can hide individual errors. Missing shared evidence is not a zero error.

View the field charts ↗
BGC87 · 7 planned tasks · Scoped computational comparison

Nuclear science

Advanced reactor inventory & decay heat; Radiation damage in structural materials; Medical isotope production; Hydrogen isotope separation; Actinide chemistry & reprocessing

Recorded delivery time

LINUR
0.174 s
Opus
375.69 s
Sol
210.69 s
Astra
243.83 s

5 matched timed tasks. Median seconds on shared complete-output cases.

LINUR: child dispatch-to-return, excluding wrapper preparation and packaging. AI: task receipt to answer-file finalization, including tools. Different boundaries—not like-for-like or hardware-normalized speed.

Reference agreement and scope

Linur: 8.57165e-16 relative-difference ratio.

Opus: 5.63438e-16 relative-difference ratio.

Sol: 7.1396e-16 relative-difference ratio.

Astra: 7.1396e-16 relative-difference ratio.

5 cases / 36 matched scalar outputs.

Relative difference is not an accuracy percentage. Scalar summaries are medians within, then across cases; medians can hide individual errors. Missing shared evidence is not a zero error.

View the field charts ↗
BGC87 · 6 planned tasks · Scoped computational comparison

Fusion

Core plasma turbulence & confinement; Plasma–material interactions; Tritium breeding blanket; High-temperature superconducting magnets

Recorded delivery time

LINUR
0.146 s
Opus
275.70 s
Sol
144.48 s
Astra
183.15 s

1 matched timed tasks. Median seconds on shared complete-output cases.

LINUR: child dispatch-to-return, excluding wrapper preparation and packaging. AI: task receipt to answer-file finalization, including tools. Different boundaries—not like-for-like or hardware-normalized speed.

Reference agreement and scope

Linur: 0.0278143 relative-difference ratio.

Opus: 0.00762203 relative-difference ratio.

Sol: 0.00602232 relative-difference ratio.

Astra: 0.00878275 relative-difference ratio.

4 cases / 22 matched scalar outputs.

Relative difference is not an accuracy percentage. Scalar summaries are medians within, then across cases; medians can hide individual errors. Missing shared evidence is not a zero error.

View the field charts ↗
BGC87 · 3 planned tasks · Scoped computational comparison

Quantum tech

Superconducting qubit decoherence; Colour-centre spin qubits

Recorded delivery time

LINUR
0.215 s
Opus
262.82 s
Sol
295.79 s
Astra
354.80 s

1 matched timed tasks. Median seconds on shared complete-output cases.

LINUR: child dispatch-to-return, excluding wrapper preparation and packaging. AI: task receipt to answer-file finalization, including tools. Different boundaries—not like-for-like or hardware-normalized speed.

Reference agreement and scope

Linur: 3.04611e-15 relative-difference ratio.

Opus: 2.07572e-16 relative-difference ratio.

Sol: 1.79658e-16 relative-difference ratio.

Astra: 1.79658e-16 relative-difference ratio.

3 cases / 27 matched scalar outputs.

Relative difference is not an accuracy percentage. Scalar summaries are medians within, then across cases; medians can hide individual errors. Missing shared evidence is not a zero error.

View the field charts ↗
BGC87 · 11 planned tasks · Scoped computational comparison

AI hardware

Post-silicon transistors; Non-volatile memory; Thermal management in 3D integrated circuits; Neuromorphic computing; EUV lithography & patterning; Wide-bandgap power electronics

Recorded delivery time

LINUR
0.167 s
Opus
637.19 s
Sol
201.08 s
Astra
294.91 s

6 matched timed tasks. Median seconds on shared complete-output cases.

LINUR: child dispatch-to-return, excluding wrapper preparation and packaging. AI: task receipt to answer-file finalization, including tools. Different boundaries—not like-for-like or hardware-normalized speed.

Reference agreement and scope

Linur: 0.0064094 relative-difference ratio.

Opus: 0.00521392 relative-difference ratio.

Sol: 0.0064094 relative-difference ratio.

Astra: 0.0064094 relative-difference ratio.

10 cases / 92 matched scalar outputs.

Relative difference is not an accuracy percentage. Scalar summaries are medians within, then across cases; medians can hide individual errors. Missing shared evidence is not a zero error.

View the field charts ↗
BGC87 · 5 planned tasks · Scoped computational comparison

AI discovery

AI-driven materials discovery; AI-driven protein design; Autonomous closed-loop experimentation

Recorded delivery time

No shared four-system timing cohort. Individual observations remain in the technical annex.

LINUR: child dispatch-to-return, excluding wrapper preparation and packaging. AI: task receipt to answer-file finalization, including tools. Different boundaries—not like-for-like or hardware-normalized speed.

Reference agreement and scope

Produced / requested outputs:

Details in the Annex: case-specific assessments, not missing calculations.

Relative difference is not an accuracy percentage. Scalar summaries are medians within, then across cases; medians can hide individual errors. Missing shared evidence is not a zero error.

View the field charts ↗
BGC87 · 5 planned tasks · Scoped computational comparison

Engineering

High-entropy alloys; Additive manufacturing; Corrosion & passivation; Rare-earth-lean permanent magnets

Recorded delivery time

LINUR
0.160 s
Opus
298.91 s
Sol
160.02 s
Astra
139.12 s

2 matched timed tasks. Median seconds on shared complete-output cases.

LINUR: child dispatch-to-return, excluding wrapper preparation and packaging. AI: task receipt to answer-file finalization, including tools. Different boundaries—not like-for-like or hardware-normalized speed.

Reference agreement and scope

Linur: 0.0453782 relative-difference ratio.

Opus: 0.0490232 relative-difference ratio.

Sol: 0.06258 relative-difference ratio.

Astra: 0.0576795 relative-difference ratio.

4 cases / 65 matched scalar outputs.

Relative difference is not an accuracy percentage. Scalar summaries are medians within, then across cases; medians can hide individual errors. Missing shared evidence is not a zero error.

View the field charts ↗
BGC87 · 11 planned tasks · Scoped computational comparison

Biomedicine

Structure-based drug design; Drug resistance; Enzyme design; Antimicrobial resistance; Biopharmaceutical formulation; Nanomedicine & drug delivery

Recorded delivery time

No shared four-system timing cohort. Individual observations remain in the technical annex.

LINUR: child dispatch-to-return, excluding wrapper preparation and packaging. AI: task receipt to answer-file finalization, including tools. Different boundaries—not like-for-like or hardware-normalized speed.

Reference agreement and scope

Produced / requested outputs:

Details in the Annex: case-specific assessments, not missing calculations.

Relative difference is not an accuracy percentage. Scalar summaries are medians within, then across cases; medians can hide individual errors. Missing shared evidence is not a zero error.

View the field charts ↗
BGC87 · 7 planned tasks · Scoped computational comparison

Chemistry

Homogeneous catalysis; Combustion & sustainable fuels; Excited-state dynamics; Spectroscopic prediction

Recorded delivery time

LINUR
0.166 s
Opus
282.63 s
Sol
159.16 s
Astra
220.06 s

2 matched timed tasks. Median seconds on shared complete-output cases.

LINUR: child dispatch-to-return, excluding wrapper preparation and packaging. AI: task receipt to answer-file finalization, including tools. Different boundaries—not like-for-like or hardware-normalized speed.

Reference agreement and scope

Linur: 0.0375277 relative-difference ratio.

Opus: 0.00300449 relative-difference ratio.

Sol: 0.00148485 relative-difference ratio.

Astra: 0.00783832 relative-difference ratio.

6 cases / 23 matched scalar outputs.

Relative difference is not an accuracy percentage. Scalar summaries are medians within, then across cases; medians can hide individual errors. Missing shared evidence is not a zero error.

View the field charts ↗
BGC87 · 5 planned tasks · Scoped computational comparison

Crystal discovery

High-temperature superconductors; Gas separation & carbon capture; Metastable phase discovery

Recorded delivery time

LINUR
0.168 s
Opus
339.19 s
Sol
178.59 s
Astra
116.49 s

2 matched timed tasks. Median seconds on shared complete-output cases.

LINUR: child dispatch-to-return, excluding wrapper preparation and packaging. AI: task receipt to answer-file finalization, including tools. Different boundaries—not like-for-like or hardware-normalized speed.

Reference agreement and scope

Linur: 0 relative-difference ratio.

Opus: 6.82753e-10 relative-difference ratio.

Sol: 8.7814e-16 relative-difference ratio.

Astra: 0 relative-difference ratio.

2 cases / 26 matched scalar outputs.

Relative difference is not an accuracy percentage. Scalar summaries are medians within, then across cases; medians can hide individual errors. Missing shared evidence is not a zero error.

View the field charts ↗

Compare observed repeat agreement across fields ↗

Request detailed technical annex under NDA →

Earlier pilot · 8 September 2026 · public report v6

The five-case pilot. Preserved.

Methods, results, timing conditions and limitations. Five scenarios; 77 scored checks per respondent. LINUR, Claude Opus 5 and GPT-5.6 Sol each passed all 77 checks.

This is scoped computational evidence from a protocol-qualification pilot, not independent experimental validation or a claim of universal capability.

Timing boundary. LINUR reports prepared-local execution; the AI comparisons report end-to-end sessions including orchestration. Preparation was completed beforehand for LINUR. These are not like-for-like or hardware-normalized speed measurements.

Prepared 8 September 2026 · 8 pages · PDF

Read the full public report (v6) ↗

Download the PDF

Explore the 14 earlier application studies

Earlier application studies · retained evidence

The earlier studies, in context.

Discipline is a way to navigate, not a claim that every subject is solved. The 87-scenario campaign and 14 application studies remain separate evidence collections.

14 of 14 unique application studies · primary discipline only

These earlier studies retain their original timing and result boundaries. The newer BGC87 field comparisons appear above; they do not automatically replace a different scientific task.

Electrical & Optical Systems

A04 · Application studies · Validated result

Passive wire antenna

Can a frozen passive-antenna operator agree with an independent solver on held-out cases?

Against independent NEC2++: 0.799 dB mean S11-curve RMSE, 10.42 MHz resonance MAE, 0.0269 dB forward-gain MAE and 0.286 dB front/back MAE.

Limit. Zero target-specific fits. This is a computational cross-solver comparison—not antenna hardware validation.

Scope and record details
Status
Validated result
Method / result classification
Held-out computational comparison
Public capability
Passive wire-antenna observables under a fixed computational comparison.
Test date / version
Not supplied as structured metadata in the catalogue; consult the approved report where available.
Timing scope
No matched complete-response timing supplied in this public catalogue. Existing report timings retain their original scope.
Cross-disciplinary tags
None
A05 · Application studies · Negative result

Steerable antenna array

Does the passive operator transfer to active beam steering?

32 cases: beam error 19.25° versus 0.25° in NEC; pattern RMSE 6.32 dB. Only the front/back gate passed.

Limit. No array or hardware claim. The result is retained as an explicit missing-operator and problem-definition finding.

Scope and record details
Status
Negative result
Method / result classification
Held-out computational comparison
Public capability
A held-out capability check against an independent NEC reference.
Test date / version
Not supplied as structured metadata in the catalogue; consult the approved report where available.
Timing scope
No matched complete-response timing supplied in this public catalogue. Existing report timings retain their original scope.
Cross-disciplinary tags
None
A06 · Application studies · Reduced interpretable result

Wireless power transfer

Can a reduced optimization identify promising geometry before a higher-fidelity model or physical test?

Median reduced-model efficiency rose from 2.52% to 6.16%, a median relative uplift of 77.8%.

Limit. Research result only. Segment convergence is insufficient for an absolute continuum-efficiency claim; no skin/proximity, ferrite, shield, tissue, thermal, EMC, power-electronics or hardware authority.

Scope and record details
Status
Reduced interpretable result
Method / result classification
Reduced numerical optimization
Public capability
A reduced numerical comparison across 120 declared cases.
Test date / version
Not supplied as structured metadata in the catalogue; consult the approved report where available.
Timing scope
No matched complete-response timing supplied in this public catalogue. Existing report timings retain their original scope.
Cross-disciplinary tags
Energy & Thermal Systems
A08 · Application studies · Validated result

Phase-shifted DFB laser

Can a locked model reconstruct a published linear threshold without target-specific fitting?

QWS threshold 19.81556 mA versus the published 19.78 mA: 0.1798% relative difference.

Limit. Historical reconstruction—not a new blind experiment. No linewidth, thermal, spontaneous-emission, above-threshold power, fabrication or hardware authority.

Scope and record details
Status
Validated result
Method / result classification
Locked historical reconstruction
Public capability
A certified linear-threshold reconstruction under declared prior knowledge.
Test date / version
Not supplied as structured metadata in the catalogue; consult the approved report where available.
Timing scope
No matched complete-response timing supplied in this public catalogue. Existing report timings retain their original scope.
Cross-disciplinary tags
Materials & Solid-State Physics

Materials & Solid-State Physics

A01 · Application studies · Reduced interpretable result

Graphene / Floquet materials

Can a large conventional lattice calculation remain governed across a declared case set?

375 exact-grid cases: 276 certified and 99 failed closed; largest admitted lattice 99,900 sites.

Limit. Conventional Dirac/Floquet/Haldane physics under LINUR certification—not experimental validation, a new microscopic law, a quantum-computer replacement or a comparative speed claim.

Scope and record details
Status
Reduced interpretable result
Method / result classification
Certified conventional-model computation
Public capability
Exact-grid Dirac/Floquet/Haldane cases under LINUR certification.
Test date / version
Not supplied as structured metadata in the catalogue; consult the approved report where available.
Timing scope
No matched complete-response timing supplied in this public catalogue. Existing report timings retain their original scope.
Cross-disciplinary tags
Electrical & Optical Systems

Chemistry & Molecular Systems

A07 · Application studies · Extension in progress

CO₂ photoreduction

Is the required physical operator available before target data are opened?

Execution stopped before target data were opened because required adsorption, reaction-coordinate, solvent, photochemical-transition, kinetic, selectivity, yield and stability operators were unavailable.

Limit. No CO₂ conversion claim. This is a transparent no-go result, not a successful application.

Scope and record details
Status
Extension in progress
Method / result classification
Capability audit
Public capability
A source-level capability audit that stops when the lawful physics is absent.
Test date / version
Not supplied as structured metadata in the catalogue; consult the approved report where available.
Timing scope
No matched complete-response timing supplied in this public catalogue. Existing report timings retain their original scope.
Cross-disciplinary tags
Energy & Thermal Systems
A09 · Application studies · Reduced interpretable result

Six-anchor electrochemical impedance

Does a six-anchor reduced spectrum reconstruction meet the declared application gate?

Best admitted lane reached median held-out complex NRMSE 0.3196; the stricter lane rejected all 519 spectra.

Limit. Six measured anchors per spectrum are required. No measurement-free chemistry, microscopic uniqueness or pH authority.

Scope and record details
Status
Reduced interpretable result
Method / result classification
Reduced calibrated reconstruction
Public capability
A calibrated model-class and anchor-sufficiency test.
Test date / version
Not supplied as structured metadata in the catalogue; consult the approved report where available.
Timing scope
No matched complete-response timing supplied in this public catalogue. Existing report timings retain their original scope.
Cross-disciplinary tags
Electrical & Optical Systems
A10 · Application studies · Validated result

Calibrated pH readout

Can a compact impedance readout transfer to closed pH labels after a declared calibration?

Five closed pH labels after four-record calibration: MAE 0.1199 pH and RMSE 0.1606 pH.

Limit. pH keeps its standard definition. This is not a proton-activity derivation, universal chemistry model or first-principles pH result.

Scope and record details
Status
Validated result
Method / result classification
Calibrated prediction
Public capability
A calibrated impedance-descriptor transfer under admitted sample and instrument conditions.
Test date / version
Not supplied as structured metadata in the catalogue; consult the approved report where available.
Timing scope
No matched complete-response timing supplied in this public catalogue. Existing report timings retain their original scope.
Cross-disciplinary tags
None
A14 · Application studies · Validated result

Molecular coordination topology

Can a frozen local-topology operator transfer across newly opened rubredoxin structures without target-specific fitting?

All 16 eligible external Zn–S₄ structures passed the frozen angular and normalized-topology panel limits.

Limit. Local coordination topology only—not a complete protein fold, absolute bond length or X-ray electron-density reconstruction.

Scope and record details
Status
Validated result · Engine-native result
Method / result classification
Derived normalized geometry · coordinate-blind external transfer
Public capability
From connectivity alone, LINUR derives the regular four-coordinate topology and transfers it across newly opened rubredoxin structures.
Test date / version
Not supplied as structured metadata in the catalogue; consult the approved report where available.
Timing scope
No matched complete-response timing supplied in this public catalogue. Existing report timings retain their original scope.
Cross-disciplinary tags
None

16 / 16 · external structures passed

4.97° · median angular MAE

0 · target-specific fits

< 1 s · forward evaluation on the local Mac

Technical evidence boundary
  • Frozen ideal angle: 109.4712206°. Sixteen of sixteen external Zn–S₄ structures were eligible and passed.
  • Median angular MAE: 4.9680°; range 4.4316° to 5.3013°.
  • Median normalized topology energy: 0.04808 against a frozen 0.15 panel limit.
  • Earlier Fe–S₄ comparisons remain preserved: 5.8312° MAE for 9TA4 and 5.6825° for 9TA6.

A general UFF radial adapter estimated the Fe–S bond scale with 0.0195 Å MAE in the earlier comparison. This is a reduced empirical result, not yet an Engine-derived absolute length.

The earlier whole-fold model did not pass its conjunction gate. A single-drive elastic network improved the 8 Å dose-response RMS by only 1.52%; the 7 Å and 9 Å variants were worse than zero response. These remain negative or weak technical results, not success metrics.

The 0.43 Å rubredoxin study identifies Fe–S₄ electronic structure as an unresolved modelling problem. LINUR’s frozen, coordinate-blind 30OH/30OR comparison remains pending because deposited coordinates were not publicly retrievable at freeze. No Max Planck electron-density result is claimed.

Read the approved report ↗

Separate conditional reconstruction and regression metrics ↓

Energy & Thermal Systems

A02 · Application studies · Reduced interpretable result

Battery charging

Which declared charging protocol reduces simulated charge time while respecting the stated thermal constraint?

Eight held-out discharges: 68.86 mV voltage RMSE and 1.55 °C temperature RMSE. Selected protocol reduced simulated charge time 17.47% under the declared 45 °C constraint.

Limit. Conventional DFN/P2D model only; no new battery microscopic physics, hardware, charging-safety or lifetime authority.

Scope and record details
Status
Reduced interpretable result
Method / result classification
Calibrated reconstruction
Public capability
Protocol screening in a conventional PyBaMM DFN/P2D battery model.
Test date / version
Not supplied as structured metadata in the catalogue; consult the approved report where available.
Timing scope
No matched complete-response timing supplied in this public catalogue. Existing report timings retain their original scope.
Cross-disciplinary tags
Chemistry & Molecular Systems
A03 · Application studies · Negative result

Ni-63 betavoltaic

Can a measurement-free device calculation recover observed output?

The measurement-free device prediction overestimated measured output by 8.018×.

Limit. No admitted physical device-performance or full-3D transport claim. An earlier exact match depended on one calibrated coefficient and is not presented as a prediction.

Scope and record details
Status
Negative result
Method / result classification
Physical-prediction attempt
Public capability
A capability test that preserves the negative result instead of fitting it away.
Test date / version
Not supplied as structured metadata in the catalogue; consult the approved report where available.
Timing scope
No matched complete-response timing supplied in this public catalogue. Existing report timings retain their original scope.
Cross-disciplinary tags
Nuclear & Isotope Studies · Materials & Solid-State Physics
A11 · Application studies · Reduced interpretable result

Eight-port ECRH control

Can an inverse controller expose an infeasible request before returning the largest admissible action?

A 10% request is infeasible inside 0.8% symmetric headroom; exact global saturation returns 0.8%. A declared 20% authority envelope delivers the full 10% request.

Limit. Reduced open-loop result only. No plasma deposition, MHD, fast-particle, divertor, hardware, delay/noise or closed-loop robustness validation.

Scope and record details
Status
Reduced interpretable result
Method / result classification
Reduced open-loop control
Public capability
Headroom-aware global scaling in a declared reduced open-loop control problem.
Test date / version
Not supplied as structured metadata in the catalogue; consult the approved report where available.
Timing scope
No matched complete-response timing supplied in this public catalogue. Existing report timings retain their original scope.
Cross-disciplinary tags
Electrical & Optical Systems

Nuclear & Isotope Studies

A12 · Application studies · Reduced interpretable result

3,030-nuclide isotope table

Can a frozen reduced model compare mass, stability, lifetime and decay mode across a large admitted table?

Binding MAE 58.92 keV/A; stability AUC 0.9041; half-life MAE 2.4455 log₁₀ s; decay-mode accuracy 83.24% on 2,762 labels.

Limit. Mixed result. The zero-fit physical mass, stability, Q-value and lifetime lane remains unavailable; no claim of predicting isotope lifetimes from geometry is made.

Scope and record details
Status
Reduced interpretable result
Method / result classification
Calibrated reduced prediction
Public capability
Element-blocked calibrated readouts across 3,030 nuclides and 92 elements.
Test date / version
Not supplied as structured metadata in the catalogue; consult the approved report where available.
Timing scope
No matched complete-response timing supplied in this public catalogue. Existing report timings retain their original scope.
Cross-disciplinary tags
None

Gravity & Cosmology

A13 · Application studies · Reduced interpretable result

Galaxy, cluster and expansion data

Can one frozen reduced comparison be audited across galaxy, cluster and expansion cohorts?

Mixed science: conventional RAR slightly beat the closure-shaped mapping in both radial cohorts, and the expansion deformation did not beat flat ΛCDM.

Limit. No new-gravity claim. Weak lensing, BAO/SN distance observables, strong-field and black-hole predictions remain unfinished.

Scope and record details
Status
Reduced interpretable result
Method / result classification
Calibrated reduced comparison
Public capability
One-scale calibrated mappings across 175 galaxies, 47 clusters and 32 chronometers.
Test date / version
Not supplied as structured metadata in the catalogue; consult the approved report where available.
Timing scope
No matched complete-response timing supplied in this public catalogue. Existing report timings retain their original scope.
Cross-disciplinary tags
None

Bring your own benchmark.

Tell us the decision you need to make, your current workflow and the accuracy that matters. We can arrange an NDA, agree a comparison protocol and assess a scoped trial.

Discuss a confidential trial ↗

Start with a non-confidential description. We can arrange an NDA before exchanging sensitive data.

Public boundary

Evidence status

Validated result

Each public statement is labelled by what it establishes—not by a generic pass or fail.

Validated, reduced, Engine-native, extending and negative results remain distinct.
Laboratory boundary

Immutable first

Engine-native result

Measurements and provenance enter the evidence registry before they can affect a model.

Frozen gaps become predeclared experiments, then target-blind improvements and independent transfer tests.
Numerical boundary

Certificate ≠ physics

Reduced interpretable result

A numerical or algebraic certificate is not a physical-accuracy claim.

Physical error and hardware authority require their own separately reported evidence.

Matter & Life · validated local topology

A local result with a deliberately narrow claim.

From connectivity alone, LINUR derives regular four-coordinate topology and transfers it across newly opened rubredoxin structures. This does not establish a complete protein fold or an X-ray electron-density reconstruction.

Validated resultEngine-native result
External structures
16 / 16
Median angular MAE
4.97°
Target-specific fits
0
Open the technical evidence boundary

Reduced scale. A general UFF adapter estimated the Fe–S bond scale with 0.0195 Å MAE. It is a reduced empirical result, not an Engine-derived absolute length.

Negative whole-fold result. The earlier whole-fold model did not pass its conjunction gate.

Weak dose response. The single-drive 8 Å network improved zero-response RMS by only 1.52%; the 7 Å and 9 Å variants were worse than zero response.

Pending continuation. The frozen 30OH/30OR local comparison remains pending. LINUR does not claim the Max Planck electron-density result.

Read the non-confidential report ↗

Matter & Life · A14 regression evidence

Conditional Fe–S₄ Molecular Reconstruction

Reduced interpretable resultExtension in progress

CALIBRATED CONDITIONAL RECONSTRUCTION / REGRESSION COMPARISON

Comparison of the LINUR conditional Fe–S₄ reconstruction with the 9TA4 and 9TA6 regression references
Regression comparison · scale and topology frozen before scoring · not new blind validation

Given Fe–S₄ connectivity, the sealed tetrahedral topology and one independently published ferric-tetrathiolate scale, LINUR reconstructed the dominant local geometry with zero target-specific fits. Across two regression structures, Fe–S bond-length MAE was 0.0218–0.0330 Å. The stricter complete-coordinate gate did not yet pass, identifying the protein-environment physics still required.

TopologyPASS

Connectivity-only sealed tetrahedral topology · 109.4712206° · zero target-specific fits.

Conditional bond scalePASS

One independent ferric-tetrathiolate scale · 2.284 ± 0.002 Å · frozen before scoring.

Complete-coordinate gateFAIL

9TA4 0.1684 Å · 9TA6 0.1661 Å aligned local RMSD · frozen gate 0.15 Å.

StructureBond MAEMaximum errorLocal RMSD
9TA40.0218 Å0.0364 Å0.1684 Å
9TA60.0330 Å0.0688 Å0.1661 Å

Numerical certificate. 1.13 × 10⁻¹⁶ relative build-bond residual. This is a coordinate numerical certificate, not a physical-accuracy claim.

Next sealed gap. R5BR-XW sealed the next missing layer: a protein-environment Hamiltonian and bounded continuous solver. This is the next target-blind development layer before independent transfer.

Evidence boundary. The scale is calibrated, and 9TA4/9TA6 are previously opened regression structures—not new blind targets. This does not establish measurement-free absolute scale, exact complete structures, a full protein fold or replacement of X-ray crystallography.

Read the evidence report ↗Gebhard et al., JACS 112 (1990) 2217–2231 ↗

How comparisons are made

A matched response-time test starts when a case is released and ends when the complete required response is received. Preparation, assistance, retries and caching rules are declared. Accuracy and completion accompany the time, with all reported trials retained. Measurement and internal implementation are separate: a black-box comparison can protect the method without concealing what the timer includes.

Status language

Explicit badges. No silent promotion.

Validated resultReduced interpretable resultEngine-native resultExtension in progressNegative result

Audit the boundary

Bring a decision and the evidence required to make it.

Challenge LINUR